Download pdf - Advanced Carbon Materials and Technology (Tiwari/Advanced) || Engineering Behavior of Ash Fills

419

Ashutosh Tiwari and S.K. Shukla (eds.) Advanced Carbon Materials and Technology, (419–474) 2014 © Scrivener Publishing LLC

11

Engineering Behavior of Ash Fills

Ashutosh Trivedi

Professor, Department of Civil Engineering, Delhi Technological University, Delhi, India

Abstract The coal-based thermal power plants often dispose of surplus fl y ash or coal ash in a nearby low-lying area normally called ash pond. It is exten-sively used as a structural fi ll and as a subgrade for highway embankment. The bearing capacity and settlement of cemented and uncemented ash fi lls are based upon a quality designation related to RQD and penetration tests, respectively. The cemented ash fi ll is characterized by a hardening parameter obtained from the joint parameters of a fractured core. Many of these fi lls in uncemented and uncompacted state show disproportionate settlement due to collapse, piping, erosion and liquefaction. For the use of these fi lls as foundation soil, the ash should be characterized and com-pacted. In the present chapter the characterization, hardening, bearing capacity and settlement are analyzed for the ash fi lls. It is based on rela-tive density and relative compaction for uncemented ash fi lls. The bearing capacity of compacted ash is a function of relative dilatancy. A plot for settlement and foundation size is utilized to obtain the settlement of com-pacted ash. The critical values of penetration resistance of standard cone and split spoon sampler in saturated condition are adjudged vulnerable to excessive settlement as shown by the collapse and liquefaction resistance. It is shown that for classifi ed gradation of ashes above a critical degree of compaction and degree of saturation, the ash fi ll may settle less than the allowable settlements.

Keywords: Ash fi lls, compaction, RQD, load tests, bearing capacity, settlement

*Corresponding author: [email protected]

420 Advanced Carbon Materials and Technology

11.1 Background

The burning of coal has been a main source of power generation for ages, but the mass production of power by thermoelectric plants has brought a mounting problem of ash disposal in the past few decades. The ash produced by the coal-fi red power plants consists of fl y ash composed of particle sizes normally less than 300 μm and bottom ash made up of signifi cantly coarser particles. The mixture of fl y ash and bottom ash is eventually disposed of in a slurry con-tainment facility known as ash pond [1]. The combined quantum of ashes, namely fl y ash, bottom ash and pond ash, are commonly called coal ash in North American countries. According to one esti-mate, the world coal consumption of 7.5 thousand million tons in the year 2011 will be closer to 13 thousand million tons in the year 2030 with proportional generation of coal ash. The principal carbon compositions in a variety of coal types, namely anthracite, bituminous, and lignite coal, vary signifi cantly. Consequently, they have ash content ranging from less than twenty, ten and fi ve per-cent, respectively. On an average, it may require a capacity build-ing to accommodate nearly a hundred to thousand million tons of ash annually. Worldwide, more than 65% of fl y ash produced from coal power stations is disposed of in landfi lls and ash ponds [2, 3]. Therefore, the behavior of ash fi ll has remained a matter of great interest equally among engineers, scientists, planners and develop-ers, contractors, and owners in the past three decades.

11.1.1 Physico-Chemical Characterization

Normally, the composite ash collected from electrostatic precipita-tors and the bottom of the hoppers of thermal power plants may be classifi ed as coal ash. The coarse ash collected from the furnace bot-tom is known as bottom ash. It is around 20 to 25% of the total ash produced. The ash is disposed of in a pond by mixing it with water to form slurry. The slurry usually contains 20% solids by weight. This method of ash disposal is called wet method. There are sev-eral environmental issues associated with ash fi ll [4–7]. The landfi ll of ash may be used as a construction fi ll if the suitable ashes are properly characterized. The fi ne ashes may collapse upon wetting. To avoid excessive settlement upon wetting, suitability of coal ash should be examined as per the criteria of collapse [8, 9] and lique-faction [10, 11]. The chemical and physical characteristics of the ash

Engineering Behavior of Ash Fills 421

produced depend upon the quality of coal used, the performance of wash-units, effi ciency of the furnace and several other factors. The physical and chemical properties of ash are infl uenced by the type and source of coal, method and degree of coal preparation, cleaning and pulverization, type and operation of power generation unit, ash collection, handling and storage methods, etc. Ash properties may vary due to changes in boiler load. The choice of furnace type such as stoker fi red, cyclone type or pulverized coal furnace is also known to affect the properties of the ash collected.

11.1.2 Engineering Characteristics

Since coal-based power is one of the most reliable means of power generation throughout the world, the continuing practice through-out various countries has been to consider it acceptable to dispose of ash in landfi lls in areas previously considered wastelands or embankments. The recent studies [12, 5, 13] indicate potential vul-nerability of the ash fi lls to failures [8, 14, 15]. However, land recla-mation using coal ash has been viably investigated in various parts of the world. Moreover, it has been a matter of interest among the scientifi c community to explore the utilization of coal ash [16–18]. During the combustion of coal, minerals are transformed to mull-ite, magnetite, tridymite, glass, etc., thus forming a composite ash. The main chemical components of coal ash are silica, alumina, iron oxide and other alkalis.

The mineral group present in coal, such as hydrated silicate group, carbonate group, sulphate group and their varying compo-sitions play a major role in determining the chemical composition of ash. As per the source of coal used in different thermal plants, the chemical composition [19] and engineering parameter [20–22] of the ashes are different (Tables 11.1–11.5) compared to natural geo-materials. The design parameters for ash as structural and embank-ment fi ll provided in IRC: SP-58: 2001 are shown in Table 11.3.

The ASTM classifi cations of coal ash are related to the percentage of calcium oxide in ash. The ashes with a high amount of calcium oxide show self-hardening pozzolanic properties in the presence of water. The pozzolanic properties of fl y ash have been documented by Mehta and Monterio [23]. Such ashes are designated as class C ash. A typical class C ash is obtained from the burning of lignite coal. The ashes from bituminous coal that do not possess self-hard-ening properties are called class F ash. The large quantum of ash

422 Advanced Carbon Materials and TechnologyTa

ble

11.

1 A

sh ty

pe a

nd c

oal q

ualit

y, c

olle

ctio

n st

age,

che

mic

al a

nd p

arti

cle

char

acte

rist

ics.

Coa

l Q

ual

ity

Sta

ge

Ash

Ty

pe

Ch

emic

al, M

iner

alog

ical

C

omp

osit

ion

Par

ticl

e C

har

acte

rist

ics

Ant

hrac

ite

bitu

min

ous

and

sub

bi

tum

inou

s co

al

ESP

ho

pper

Low

ca

lciu

m

oxid

e fl

y as

h

Mos

tly

silic

ate

glas

s co

ntai

ning

al

umin

um, i

ron,

and

alk

alis

. Sm

all a

mou

nt o

f cry

stal

line

mat

ter

pres

ent g

ener

ally

con

sist

s of

qua

rtz,

mul

lite,

sill

iman

ite,

he

mat

ite,

and

mag

neti

te [2

3].

15 to

30

perc

ent p

arti

cles

larg

er

than

45μ

m (s

urfa

ce a

rea

20 to

30

m2 N

-1) m

ost o

f the

par

ticl

es

are

solid

sph

eres

20μ

m a

vera

ge

dia

met

er. C

enos

pher

e, p

lero

-sp

here

s m

ay b

e pr

esen

t [23

].

Lig

nite

coa

lE

SP

hopp

erH

igh

calc

ium

ox

ide

fl y

ash/

ty

pica

l A

STM

C

-618

cl

ass

C

ash

Mos

tly

silic

ate

glas

s co

ntai

ning

ca

lciu

m m

agne

sium

, alu

min

um,

and

alk

alis

. Sm

all a

mou

nt

of c

ryst

allin

e m

atte

r pr

esen

t ge

nera

lly c

onsi

sts

of q

uart

z,

tric

alci

um a

lum

inat

es; f

ree

lime

and

per

icla

se m

ay b

e pr

esen

t. U

nbur

ned

coa

l is

usua

lly le

ss

than

2 p

erce

nt [2

3].

10 to

15

perc

ent p

arti

cles

larg

er

than

45μ

m (s

urfa

ce a

rea

30 to

40

m2 N

-1) m

ost o

f the

par

ticl

es

are

solid

sph

eres

less

than

20

μm in

dia

met

er. P

arti

cle

surf

ace

is g

ener

ally

sm

ooth

bu

t not

as

clea

n as

low

cal

cium

ox

ide

fl y

ash

[23]

.

Any

coa

l qu

alit

yFu

rnac

e bo

ttom

Bot

tom

ash

Res

idue

con

sist

ing

of s

ilica

, mul

lite,

si

llim

anit

e, h

emat

ite,

and

m

agne

tite

[1].

Sand

siz

e pa

rtic

les

of r

ough

te

xtur

e [1

].

Any

coa

l qu

alit

yA

sh

pond

or

ash

la

goon

Pond

ash

Pond

ash

is a

mix

ture

of E

SP a

sh a

nd fu

rnac

e bo

ttom

ash

wit

h w

ater

. It

con

tain

s a

wid

e ra

nge

of p

arti

cle

size

s, te

xtur

e an

d c

hem

ical

co

mpo

siti

on d

epen

din

g up

on th

e qu

alit

y of

the

char

ge. I

t has

som

e ag

ing

affe

cts

also

[1].

Engineering Behavior of Ash Fills 423Ta

ble

11.

2 C

hem

ical

com

posi

tion

of c

oal a

shes

[1, 8

, 24]

.

Ch

emic

al c

omp

osit

ion

%B

riti

sh a

sh

Am

eric

an a

shS

wed

ish

ash

P

olis

h a

shIn

dia

n a

sh

Si O

238

–58

30–5

830

–53

43–5

248

.4–5

7.5

Al 2 O

320

–40

7–38

14–3

319

–34

18.2

–27.

2

Fe2 O

36–

1610

–42

10–1

40.

7–10

.711

.3–5

.4

Ca

O2–

100–

130.

9–6.

11.

7–9.

411

.8–3

.1

Mg

O1–

3.5

0–3

4–6

1–2.

93.

6–0.

4

Na 2O

, K2O

2–5.

50.

4–2

1.6–

3.5

0.4–

0.9

0–0.

9

S O

30.

5–2.

50.

2–1

0.4–

1.5

0.3–

0.8

0–2.

9

Unb

urne

d C

arbo

n–

0–4.

80.

9–3.

31.

9–9.

91.

2–4.

1

424 Advanced Carbon Materials and TechnologyTa

ble

11.

3 E

ngin

eeri

ng p

aram

eter

s fo

r co

al a

sh [3

, 9].

Par

amet

erIR

CU

nce

men

ted

Ash

Cem

ente

d A

sh

Sp. g

ravi

ty

Plas

tici

ty

Max

imum

dry

uni

t wei

ght

OM

C

1.90

–2.5

5 N

P9–

1618

–38%

1.70

–2.6

P,

NP

8–18

20–4

0%

2.0–

2.6

– 10–1

820

–40%

Coh

esio

nne

glig

ible

Nil

Max

of 1

MPa

Ang

le o

f int

erna

l fri

ctio

n30

–40

27–4

430

–47

Coe

ffi c

ient

of c

onso

lidat

ion

cv (c

m2 s

ec-1)

1.75×

10–5

to2.

01×

10–3

– –

– –

Com

pres

sion

ind

ex (c

c)0.

05–0

.40.

01–0

.006

–

Perm

eabi

lity

(cm

sec

-1)

8×10

–6-7×

10–4

1–

7×10

–510

–7 to

10–6

Part

icle

siz

e d

istr

ibut

ion

Gra

vel

Sand

Si

lt

Cla

y

1–10

%8–

85%

7–

90%

0–

10%

5% 10–9

0%

10–9

0%

0–5%

Inta

ct to

Fra

gmen

ted

Coe

ffi c

ient

of u

nifo

rmit

y3.

1–10

.72–

12


Table 11.4 Engineering parameters for a few uncemented geomaterials [3, 25–27].

Sand Type cu emin emax φc Reference

Monterey #0 sand

1.6 0.57 0.86 37.0 [27]

Tieino sand 1.5 0.57 0.93 34.8 [26]

Toyoura sand 1.27 0.61 0.99 35.1 [26]

Ottawa sand (round)

1.48 0.48 0.78 29.0 [26]

Sacramento river sand

1.47 0.61 1.03 33.3 [27]

Hokksund sand 1.91 0.55 0.87 36.0 [26]

Yamuna silty sand

2–35 0.31 0.78 25–30 [25]

Coal ash 2–10 0.6–0.8 1.4–2.10 27–30 [3]

Table 11.5 Peak and ultimate angles of friction for a few geomaterials [3, 25, 27, 28].

Material φ'peak φ'ult Reference

Dense well-graded sand or gravel, angular grains

55° 35° [27]

Medium dense uniform sand, round grains

40° 32°

Dense sandy silt with some clay 47° 32°

Sandy silty clay (glacial) 35° 30°

Clay-shale, on partings 35° 25°

Clay (London) 25° 15°

Yamuna silty sand 46° 25° [28]

Coal ash 50–53° 27–30° [3]


produced may be classifi ed as class F. Coal ash is normally used in the construction of ash dykes, reclamation of low-lying land, fab-ricated earth structures such as embankments, road fi lls, etc. The landfi ll’s intensive utilization of coal ash requires stability analysis of fi ll.

11.1.2.1 Characterization

X-ray diffraction study is normally carried out to identify the min-eral phases present in the ash. A typical X-ray diffraction shows that the ash contains traces of aluminum silicate, quartz and some heavy minerals like hematite and magnetite. Identifi cation of defi -nite crystalline mineral can be based on Bragg’s equation [λ = 2 d sin 2θ] where λ is wavelength of X-ray specifi c to the Cu target ele-ment [= 1.542Å] and d is interplanner spacing. Normally the test is conducted between 0°–70° (2θ), at a rate of 0.8°/sec using the CuKα characteristic radiation of Cu target element. The interplan-ner spacing of respective peaks on the x-ray pattern are calculated from the corresponding 2θ angle. These peaks are associated with the characteristic minerals. In crystalline form, ash contains traces of aluminum silicate, quartz and some heavy minerals [1].

Figure 11.1 shows an example of a typical X-ray diffraction pat-tern of an ash sample. The peak near 26.40° is characteristic of alu-mina-based silicate minerals. Their respective peaks near 33.2° and 35.4° indicate the presence of a heavy mineral like hematite or mag-netite. A subordinate amount of 11CaO·7Al2O3 is evident from peak incidentally close to 18°. The concurrence of a strong peak close to 26.5° indicates that quartz is one of the major constituents along with alumina-based silicate mineral. The crystalline silica sand is

1.0k

010 2θθ Degree 70

SiO 2

SiO 2

Fe 2 O 3

A16 Si 2 O 13

AU

Figure 11.1 X-ray diffraction pattern of a typical ash sample [1].


also characterized by this peak. The crystalline silica sand is the main mineral component of granular fi ll hydraulically deposited by rivers worldwide. The potential clay minerals may be present or absent in the ash, indicating that ash may or may not have any structural clay cohesion in its natural state. Any peak associated with the hydrated calcium silicate group that is responsible for the development of cohesion due to chemical reaction (marked by the formation of crystals of hydrated calcium aluminium silicate on cur-ing in the presence of water with time) indicates the self-hardening properties of ash. Therefore, the absence or presence of respective peak is classifi ed as cohesionless or cemented ash mass while eval-uating its behavior as an engineering fi ll. The ash samples contain divergent amounts of amorphous phase. The amorphous phase is in the highest amount in the pond among all ash types. This is because of the presence of unburned coal in bottom ash compo-nent. Comparing the X-ray diffraction pattern of ash samples with sand (cohesionless) it is implicit that sand has peaks with crystal-line quartz, while the ash has peaks of quartz as well as humps of non-crystalline matter. Burning of coal at high temperature and sudden cooling of ash in a short interval produces non-crystalline matter in coal ash. The presence of glassy phase, which is non-crys-talline in nature, is around 60 to 88% of ash by weight [23]. The cohesionless soils of similar gradation as that of ash may be charac-terized as sandy silt to silty sand. The engineering behavior of natu-ral geomaterials is also affected in a similar manner [24–28]. The cemented ash fi ll may be characterized as weakly cemented rock. These fi lls have dominant presence of crystalline quartz. The pres-ence of amorphous matter along with crystalline quartz induces signifi cant differences in the engineering characteristics of ash fi lls compared to the natural geomaterials [1].

11.1.2.2 Chemical Composition

The chemical composition of the ashes is obtained from the non-combustible components produced by burning of the coal. The comparison of a typical range of chemical composition of ashes from different parts of the world along with a typical Indian ash is given in Table 11.2. The main constituent of the ash is silica followed by alumina, oxides of iron and calcium. The presence of sodium and potassium salts is known by the qualitative chemical analysis. The submergence of ashes is critical compared to the other granular


soils due to the presence of few soluble matters. The solubility of ash sample is determined separately by boiling in water and then at room temperature. Each sample is thoroughly mixed with the boiling water by a stirrer. The entire experiment is repeated with the cold-water mix at room temperature. This mixture is fi ltered through the Whatman-42 fi lter paper. The retained ash is dried in an electric oven at 105° for 24 hours and the percentage soluble in the ash is obtained. The pond ashes have negligible soluble con-tent, while fi ne ashes obtained directly from electrostatic precipita-tor have a signifi cant percentage of soluble.

11.1.2.3 Electron Micrographs

The micrographic investigation of a typical ash sample is presented in Figure 11.2. The electron micrograph [2] indicates presence of predominantly coarse grain and fi ner particles together. It sug-gests that coarse ash contains rounded spherules, subrounded, and opaque particles. The ash contains superfi ne that form agglomer-ates which have a tendency to stick together and appear as larger particle upon pressing. Upon observing ash with a microscope it is seen that ash particles are clear or translucent spherules (siliceous aluminous particles), subrounded and rounded porous grains, irregular agglomerated glass spherule, opaque dark gray and red angular grains of magnetite and hematite, and black porous grains of carbon.

Figure 11.2 Electron micrograph of a typical ash sample [2].


11.1.2.4 Grain Size Distribution

The grain size distribution of varied ash sample indicates par-ticle size in the range of coarse sand, silt to clay. However, the maximum frequency of particle is in the range of fi ne sand to silt (Figure 11.3). The effective size and mean sizes may con-trol permeability (Figure 11.4), shear strength (Figure 11.5) and compaction (Figures 11.6, 11.7) of ashes. The pond ash which is

100

80 F7

F6

F5

F4

PA2PA1

PA1

PA2

F1

F2

F3

F4

F5

F6

F7

MH

Sand

F3F2

MH Sand

F1

60

40

20

00.001 0.01 0.1

Grain size (mm)

% F

iner

by

wei

ght

1 10

Figure 11.3 Grain size distribution of varied ash samples.

0.001

K10 = 0.00031n(D10) + 0.002R2 = 0.9558

Effective size

Mean size

1.0k SiO 2

A16 Si 2 O 13

SiO 2Fe 2 O 3

010 702θ Degree

AU

K50 = 0.00021n(D50) + 0.001R2 = 0.9102

0.0001

Per

mea

bili

ty (

mm

/s)

Effective and mean sizes (mm)

0.000010.001 0.010 0.100 1.000

Figure 11.4 Permeability characteristics of ashes.


10

9

8

7

6

5

4

3

2

1

0

–1

–2

–30.0 0.1 0.2 0.3 0.4 0.5

250 kPa

In = 12.682RD – 2.7917 R2 = 0.9628

In = 12.183RD – 2.795 R2 = 0.9603

In = 11.218RD – 2.8209 R2 = 0.9549

150 kPa50 kPaBolton (1986)Trivedi & Sud (2002)

RD

In

0.6 0.7 0.8 0.9 1.0

Figure 11.5 Shear characteristics of ashes.

22 F1F2F3F4F5F6F7S1A1ZERO VOID LINE FOR F1

ZERO VOID LINE F7ZERO VOID LINE S1ZERO VOID LINE A1

20

18

16

14

12

10

85 10 15 20 25 30

Dry

un

it w

eig

ht

(kN

/m3)

Water content (%)

35 40 45 50 55

Figure 11.6 Compaction plot for typical ash samples.

examined for mass behavior as a fi ll may contain 5 to 10% of particles in coarse and medium sand size, 35 to 50 % in fi ne sand size and 40 to 60% of particles in the range of silt. The presence of superfi ne (size ~ 0.01 mm) transforms inter-particle friction, agglomeration and formation of pendular bonds in the presence of moisture.


22 123456

78

S1A1Zero void line for G = 2.7

21

20

19

18

17

16

15

145 10 15

Water content (%)

Dry

un

it w

eig

ht

(KN

/m3)

20 25

Figure 11.7 Proctor compaction of a few natural geomaterials of similar gradation.

11.1.2.5 Apparent Specifi c Gravity

The ashes may have lower apparent specifi c gravity than the natu-ral soils of similar gradation, which is largely composed of α and β quartz, cristobalite and tridymite. The ash contains maximum percentage of silica among all the constituents. A low value of the specifi c gravity is attributed to the trapped micro-bubble of air in the ash particle and the presence of unburned carbon. The air voids percentage of ash (5 to 15%) is found to be greater than natural soils (1 to 5%) at maximum dry density [29]. It is noticed that as fi neness of the ash increases, the specifi c gravity also increases partly due to the release of entrapped gases. Investigators [30] reported a similar phenomenon in the ash grounded by mortar and pestle, indicating the possibility of the breaking of bigger particles only. The min-eralogical composition is one of the other reasons for variation in the specifi c gravity of the ash relative to soils. The ashes with high iron content tend to have a higher specifi c gravity. The presence of heavier minerals such as hematite and magnetite results in a higher specifi c gravity. It is indicated that the bottom ash typically has a


higher specifi c gravity. The pond ash may have a higher specifi c gravity than the other ashes. This is partly due to the presence of bottom ash in the pond, which contains heavier components of the coal ash. Some of the ash solids contain pores, which are not inter-connected, and hence they possess, on measurement, less specifi c gravity, although the specifi c gravity of constituent mineral remains in the usual range. Such cases are referred to as apparent specifi c gravity, which is based on the weight in air of a given volume of ash solids, which includes the isolated voids [1].

11.1.2.6 Compaction

In the design of ash containment facility, road embankments and ash fi lls, it is desirable to consider the compaction characteristics of the ashes along with natural geomaterials (Figures 11.6; 11.7). The hydraulically disposed of ash in the ash ponds is normally at a low-density state. In order to improve its engineering properties com-paction is a prerequisite. The coal ash is compacted by vibration if non-plastic in nature. However, owing to the signifi cant percent-age of fi nes, it is often compacted by impact. A granular material is found in varying states of density, i.e., loosest state to dense states.

The void ratio of ash sample in the loosest state is often obtained by a slow pouring technique. The ash is normally poured in a fi xed volume mold from a constant height of fall of 20 mm as the tech-nique applied to loose cohesionless fi ll. In the vibration test, ash is deposited at varying moisture contents in a standard thick-walled cylindrical mold with a volume of 2,830 cm3. The ash is vertically vibrated at double amplitude of 0.38 mm for seven minutes in this mold mounted on a vibration table with a frequency of 60 Hz. The diffi culties associated with fl ow of the fi nes are encountered in using this technique. The capping plate is modifi ed to fi t at the top of the mold so that it presses the ash with at least some clearance on the side. The double amplitude of vertical vibration of 0.38 mm is found to be optimum for ash samples [1].

The typical result of Proctor and vibratory compaction of the ashes with varying gradation indicates reduction in water require-ment to achieve maximum density with fi neness (Tables 11.6; 11.7; 11.8). The increasing fi neness demonstrates a sharp increase in maximum dry density in the Proctor test. Normally, the density in the vibration test is lower than that in the Proctor test on the dry side of optimum due to the rebound action of the spherical ash

Engineering Behavior of Ash Fills 433Ta

ble

11.

6 G

rain

siz

e, s

peci

fi c g

ravi

ty, P

roct

er d

ensi

ty a

nd o

ptim

um m

oist

ure

cont

ent.

Ash

Typ

eC

uD

50 (m

m)

D60

(mm

)D

10 (m

m)

Sp

Gra

vity

Max

Un

it W

eigh

t (k

N/m

3 )O

MC

(%)

Pond

8.33

0.1

0.15

0.01

81.

989.

5040

Pond

5.35

0.05

0.07

50.

014

2.00

10.3

37.5

Hop

per

3.4

0.02

30.

030.

0088

1.90

11.7

33

Tab

le 1

1.7

Res

ults

of v

ibra

tory

and

Pro

ctor

com

pact

ion.

Sp

Gra

vity

γ dm

inem

axγ d

max

(D

ry)

γ dm

ax

(moi

st)

emin

(D

ry)

emin

(moi

st)

γ dm

ax

Pro

ctor

Voi

d R

atio

at γ

dm

ax

Pro

ctor

1.98

7.63

1.6

9.56

9.50

1.06

1.08

9.5

1.08

2.00

7.85

1.54

10.5

610

.30.

890.

9410

.30.

94


Table 11.8 Gradation and compaction trends for natural geomaterials (Fig.11.7).

S.no Description Sand (%) Silt (%) Clay (%)

1 Well graded loamy sand 88 10 2

2 Well graded sandy loam 72 15 13

3 Medium graded sandy loam 73 9 18

4 Lean sandy silty clay 32 33 35

5 Lean silty clay 5 64 31

6 Loessial silt 5 85 10

7 Heavy clay 6 22 72

8 Poorly graded sand 92 6 –

9 Sand 99.3 0.63 –

10 Silty sand 96.4 3.55 –

particles at a low degree of saturation. In the vibration test a reduc-tion in the density is observed with moisture content contrary to the Proctor test. This is due to the slacking of ash at a low saturation level. The minimum value of the dry unit weight is observed at crit-ical moisture content. The dry unit weight increases beyond critical moisture due to the contravention of the surface tension force. The maximum dry unit weight is obtained at slightly higher moisture content in the vibration test.

The maximum dry unit weight of coal ash is less than that of the natural soils. This is partly due to a low specifi c gravity and a high air void content. The maximum dry unit weight by Proctor test is obtained at signifi cantly high moisture content. The maximum dry unit weight by the vibration test is slightly higher (~4%) than the Proctor test. High optimum moisture content (OMC) is normally because of the porous structure of particles. The higher OMC of coal ashes compared to natural soils is due to a large percentage of the water inside the particles in initial stages. It is diffi cult to work the particles to higher density at lower moisture. The total air in porous structure is hardly expelled to saturate ash up to OMC (Figure 11.6). Hence, the vibratory densifi cation technique resulted


in a maximum dry unit weight in dry conditions only. But, the dry state compaction is not very useful in the landfi ll. By a slight vibra-tion, ash becomes airborne and remains in the air for many hours. Sprinkling a little water helps to get rid of this problem but leads to the bulking of ash. In the standard equipment there is no fur-ther improvement in unit weight beyond 5–8 minutes of vibration. Moreover, the fi ner sample shows less variation in unit weight over time than the coarse sample. This is due to the greater inter-parti-cle friction in fi ne ashes, as found in the case of powders [31]. All ashes initially show bulking there after they reach a stage of mini-mum dry unit weight at critical moisture content. The compaction beyond optimum moisture is quite erratic. The presence of porous particle (unburned coal, pelerosphere and cenosphere) increases the optimum moisture content [1].

The ash is often compacted in the fi eld by vibratory rollers. In vibratory compaction, the maximum density in dry state is more than that in wet condition. The air voids remain entrapped among the hollow particles in wet condition. A resulting low density is attributed to surface tension in partly wet condition. The surface tension force hinders the compaction process. The maximum dry density in wet condition also coincides with the maximum dry density as obtained in the Proctor test. The modifi ed Proctor test suggests that there is no perceptible increase in the maximum dry density owing to increased compaction effort. Apparently, it is because ash is a non-plastic material [1].

The usual practice is to compact ash by vibration using a 10 to 20 ton vibratory roller at moisture content close to optimum. Tests indicate that vibratory roller compaction gives test results when soil moisture is slightly higher than optimum moisture content as obtained in the Proctor test [32]. Depending upon the proximity of available moisture with optimum moisture content, degree of com-paction may vary from 80% to 100%.

In such a limited area vibratory roller compaction was not pos-sible. Another type of vibratory compaction equipment is the vibrating base plate compactor. It produces results similar to that of vibratory rollers [32].

In fi eld compaction of ash, the use of a hand-operated base plate compactor has been reported near the foundation walls [30]. The use of heavyweight vibrators with low frequency is suggested for gravel. Light- to medium-weight vibrators with high frequency are suggested for sands and fi ner materials [32]. One of reasons for the


selection of high frequency, low weight, and base plate compactor for the ash fi ll is as sited above. Moreover, any surcharge is found to reduce the amount of densifi cation in the case of ash compacted at constant moisture content [33].

The weight and frequency of the compactor controls the thick-ness of compacted lift. The lightweight, high-frequency compactors obtain satisfactory densities in thin lifts and heavyweight, low-fre-quency compactors obtain satisfactory density in thick lifts [32]. A plate compactor of 220 N and a plate size of 152-mm x 390-mm is proposed for vibratory compaction [1]. Vibration is induced on the loose left of 150 to 200 mm. The time of vibration required is settled after several trials. Three passes are required at a frequency of 50 cps. This produces satisfactory results of density at selected moisture contents. The moisture content density data obtained by core cutters at several locations and depth on the test area is plotted along with the data from the laboratory vibration test to provide a guideline for the compaction of ash fi ll. Ash becomes airborne by slight vibration in dry state and remains suspended in air for a long time. Therefore, compaction below 5% of moisture content is not recommended.

11.1.2.7 Permeability

The permeability of coal ash can be determined at Proctor density by the falling-head method [17]. The permeability of ash depends on particle size distribution and void ratio. Pond ash deposits are loosely stratifi ed, and as a result the values of permeability are higher due to high void ratio. The permeability of ash may be related to effective size or mean sizes at Proctor density (Figure 11.4). The co-relation coeffi cient for effective sizes was found to be 0.955 compared to 0.912 for mean sizes of ashes. The permeability may be expressed [17] as a function of grain sizes at Proctor density as follows:

k = 0.0003 ln D10 + 0.002 (11.1)

k = 0.0002 ln D50 + 0.001 (11.2)

where k is permeability in mm/s, D10 is effective size and D50 is mean size in mm.

Attempts were made to relate permeability of ash samples with void ratio. At higher void ratio, it may be several times that of


permeability at Proctor density. However, in loose state, internal erosion plays a greater role than permeability. There is no defi nite trend observed between permeability and void ratio. The scatter may probably be due to the difference between actual void space available for fl ow and calculated void ratio. The calculated void ratio includes the blocked space occupied by pores of particle (porous unburnt coal, pelerosphere and cenosphere).

These values are close to the permeability of medium- to fi ne-grained soil. The permeability of medium fi ne sand and silt (SM, SL, SC) is in the range of 10–2 to 10–6 mm/sec according to the Unifi ed Soil Classifi cation System.

11.1.2.8 Compressibility

Compressibility is an important parameter to estimate the settle-ment of a ground in stressed conditions. The compressibility of ash was estimated in a 60-mm diameter and 20-mm thick oedometer ring on reconstituted loose dry samples [17]. Samples were pre-pared in an oedometer ring by the dry pluviation method. The dry ash was funnelled with zero potential energy in the ring to obtain sample in the loosest density. The sample was then pressed under a surcharge of 1kPa and temping to obtain desired specifi c volume. All reconstituted ash samples show a common preconsolidation pressure of 100 kPa, probably because of the exposure of ash mate-rial to common thermal stresses in the furnace during formation. With increasing coarseness, the compression of coal ash closely resembles sandy soils. There are certain evidences of grain crush-ing with increasing effort in tests on coarse ashes [17].

At higher consolidation pressures the compression curves of fi ne ashes may be represented by a unique normal consolidation line and their specifi c volume may be determined by the current state of stresses. There is a progressive increase in stiffness with increas-ing stresses and increasing mean sizes [17]. The stiffness is higher for samples having closeness of placement void ratio to their mini-mum void ratio. Ashes that are predominantly fi ne show a mini-mum void ratio signifi cantly lower than coarse ashes. Compression tends to pack them in a denser state and there is negligible rebound on unloading. The specifi c volume of ashes in an oedometer at varying pressures may be represented by:

V = m exp [–0.4 D50/ Da] (11.3)


Here V is the specifi c volume and D50 is the mean size of ashes and Da is a reference size parameter. The values of m depend upon the range of pressure. Typically m takes a value of 2.2 to 2.0 in a range of pressure 100 to 1600 kPa [17].

11.1.2.9 Shear Strength

The shear strength of ash in a triaxial test can be estimated at nat-ural densities and confi ning pressure for different ashes [21]. The Equations 11.4 and 11.5 show the relationship among relative dilat-ancy and relative density. Bolton [27] reviewed a large number of triaxial and plane strain cases to propose a unique relationship for dependence of frictional properties of cohesion less soil on dilation and effective mean confi ning pressure (p’) in kPa (using pa as the reference pressure in the same units). In the case of triaxial,

In= 0.33(fp – fc) + RD*ln (p’/pa) (11.4)

In= Q*RD – r (11.5)

The Equations 11.4 and 11.5 contain terms corresponding to peak (fp) and critical friction angle (fc). Here 0.33(fp - fc) is referred to as the dilatancy index in the triaxial case, with RD being relative den-sity and Q, r being material fi tting constants. Bolton [27] suggested that fi tting constant Q and critical angle are unique to a granular media that depends upon mineralogical factors. Many subsequent investigators have indicated effects of fi nes on selected parameters Q and r [26]. The value of Q for ash varied in a wide range for a selected constant value of r (Fig. 11.5). These results are compared with the value of Q (=7.7) and r (= 0) suggested by [3], ignoring low relative density values of Ir corresponding to peak frictional strength (Equation 11.6 and 11.7).

0.33(fp – fc) + RD*ln (p’/pa)= Q RD – r (11.6)

0.33(fp – fc) + RD*ln (p’/pa) = 7.7 RD (11.7)

The partly saturated ash exhibits cohesion owing to pendular bonding. The apparent cohesion is lost upon wetting and therefore should not be considered for design purposes. In the highest pos-sible packing, the angle of internal friction may be as high as 60°.


The angle of internal friction of coal ash is compared with that of sand. It may be observed that at low relative density (0–60%) its peak frictional angle is higher than clean quartz sands, while at high relative density, its frictional properties are similar to silty sand [17].

11.2 Engineering Evaluation of Cemented Ash Fill

It has been observed that aged high-lime ash behaves similarly to weakly cemented fi ll, which consists of fracture under a gradual process of discontinuous depositions and self-setting. The low-lime ash fi lls under pressure also harden to show behavior simi-lar to cemented fi ll. The vast assemblage of studies on cemented fi lls [34–37] indicate that low cemented granular materials show cohesive behavior up to a limiting strain level, and thereafter they take load by friction only. Therefore, the engineering characteristics of cemented ash fi ll can be well characterized by the techniques applied for weak and fractured rock masses [38, 39].

11.2.1 Measurement of Cemented Ash Characteristics: Application of RQD

The cemented ash fi lls consist of strong and weak formations simi-lar to the rock masses, which appear to be a continuous deposit. The drilling processes tend to recover a continuous core of strong material. The already existing weaknesses reappear as discontinu-ity at irregular frequency, which is the average number of joints appearing per unit length.

The RQD [40–42] is normally considered as the percentage of a core recovery of spacing length of greater than or equal to 100 mm. The RQD also captures the orientation of the discontinuity relative to the scan line, and thus we obtain the orientation parameter [43]. This information should be recorded in the data sheet of RQD. The engi-neering applications usually consider RQD as the percentage of the borehole core in a drill run consisting of intact lengths of rock greater than or equal to 100 mm, which is represented numerically as,

1100 %

n i n

iRQD L L

== ∑

(11.8)


where Li is the lengths of individual pieces of core in a drill run having lengths more than or equal to100 mm, and Ln is the total length of the drill run.

11.2.2 Concept of Strength Ratio and Modulus Ratio

The strength of weak rock-like-masses, namely cemented ash fi ll, is recorded in terms of strength ratio (smr). The aim of fi nding the strength relationship with modifi ed joint factor is to readily get the strength and modulus of fractured cemented ash fi ll by conducting a compression test on both intact cemented ash fi ll and concrete-forming embedded foundation. The strength ratio (smr) is defi ned as a ratio of strength of a fractured ash mass (sm) as compressive strength of least (of the weak material) of its size intact cemented ash sample (sr) or concrete sample. If s1m, s2m, and s3m are triaxial principal stresses in the jointed rock andsr is uniaxial compressive strength of concrete or intact cemented ash sample, then in the tri-axial state, the strength ratio is defi ned by:

p = (s1m +s2m +s3m)/3 (11.9)

q = [(s1m -s3m)] (11.10)

smr = [(s1m +s2m +s3m)/3]/ [(sr)/3] (11.11)

= p/ [(sr)/3]

smr = 3p/ (sr) (11.12)

In unconfi ned state the strength ratio is:

smr = [(sm)/3]/ [(sr)/3] = [sm]/ [sr] (11.13)

Similarly, confi ning pressure ratio and shear stress ratio are defi ned as:

pmr = [(s1m +s2m +s3m)/3]/ [(sr)] (11.14)

qmr = [(s1m -s3m)]/ [(sr)] (11.15)

In triaxial conditions, modulus ratio is defi ned as:

Emr = Em /Er (11.16)


Er and Em are tangent modulus of intact and fractured sample, respectively, which may be considered at hardening for sr and sm, respectively.

Similarly, in unconfi ned conditions, modulus ratio is defi ned as:

Emr = Em (s2m =s3m =0)/Er (11.17)

The relationship of strength ratio [43] and joint factor is repre-sented by:

smr = exp (ap Jfgp) (11.18)

As per a few of the investigators [38, 39, 42, 44], ap is a dilatancy-based parameter.

The variation of strength ratio with joint orientation, joint num-ber and friction shows variation of strength ratio according to the experimental observation of various investigators. Further, the modulus ratio and hardening parameter [42] are represented by:

Emr = exp (Ch Jfgh) (11.19)

Ch = κ(pi/sr)η (11.20)

The pressure dependence of hardening parameter is shown in Figure 11.8. The limiting and appropriate value of hardening

Values of αρ

[–0.005]

1

0.1

0.01

0.001

0.0001

0 5 10 15 20

Pressure ratio (%)

–Ch

25 30 35 400.00001

[–0.009][–0.01][–0.0125]

Figure 11.8 Hardening parameter Ch for embedded foundation [42].


parameter is adopted and calculated according to the end condi-tions as shown in Tables 11.5 and 11.6, respectively.

11.2.3 Evaluation of Joint Parameters

The numbers of joints per unit run of the fi ll core obtained for esti-mating RQD is joint number, Jn. Its orientation affects the equivalent number of horizontal joints per unit volume of cemented ash mass. A joint factor captures potential engineering possibilities within a joint, namely number of joints, orientation, and friction and cohe-sionless infi ll material. The condition of joints, namely presence of cohesionless infi ll material, was considered in modifi ed joint factor [44, 38]. The joint number connects RQD, joint factor and modifi ed joint factor. According to several investigators [41, 43–46], the joint factor is a signifi cant joint mapping parameter in relation to the strength ratio of rock-like masses. A concept for model behavior of cemented ash mass with joint inclination, joint number and friction considers the number of joints, joint orientation and friction in the expression as:

Jf = Jn / nb r (11.21)

where Jn is the number of joints per unit length in the direction of loading (joints per meter length of the sample); nb is orienta-tion parameter corresponding to the angle of inclination of joint (b ) to the load direction and; r is reference roughness parameter.It is presented in nondimentional form as:

Jf = Jn Lna / nb r (11.22)

where Lna is a reference length = 1 meter. On the basis of experimental data of several investigators [43],

the joint factor was modifi ed [44] to incorporate varied engineering possiblities amid cemented ash mass as:

Jfg =cg Jf (11.23)

where cg is a modifi cation factor for the condi tion of the joint, water pressure and the cohesionless infi ll material,

cg= Jdj Jt /(gd Jw ) (11.24)


where Jdj is the correction for the depth of joint (joint stress param-eter); Jt is the correction for the thickness of cohesionless infi ll mate-rial in joint (thickness parameter); Jw is the correction for ground water condition and; gd is the correction factor depending upon the compactness or relative density of cohesionless infi ll material in joint, equal to unity for fully compacted joint fi ll.

For clean compact joints, cg is equal to unity. These discontinui-ties may consist of fragments of the ash material to a varied extent of thickness, density and orientation. An observation of cemented ash core recovery captures gd in terms of designated discontinuity condition or precisely by assigning a packing density of fragments in the discontinuity. The packing in the discontinuity tends to com-pact, dilate or crush during the process of loading.

The progressive compression of the discontinuities signifi cantly infl uences the strength ratio and deformation of the cemented ash mass. The granular material in the joint’s rupture zone undergoes shear deformation depending upon relative density of the cohe-sionless infi ll material. The relative density (RD) is conventionally considered as a ratio of difference of natural state void ratio (en) and minimum void ratio (emin) from maximum void ratio (emax) of the infi ll material as:

RD= (emax– en)/ (emax– emin) (11.25)

The effect of pore pressure (u) is adjusted so that mean effec-tive confi ning pressure (p¢) is equal to mean confi ning pressure (p) [p= (s1 +s2 +s3)/3, where s1, s2, s3, are principal stresses].

11.2.4 Relationship of RQD and Joint Parameters

Several investigators [40] showed that RQD value changes with increasing difference between size, i.e., joint spacing. In fact, chang-ing block size modulates stress intensity on discontinuity, hence upon the volume of joint material. Such an observation calls for an adjustment in RQD vs joint properties, namely spacing, volume, friction, gouge material, ground water and internal pressure. The volumetric joint count tends to have an exponential relation with RQD as block size increases.

Jfg = a exp (b RQD) (11.26)


where a, b are fi tting constants, which take a value according to the rock block characteristics, discontinuity and friction in relation to RQD.

The volumetric effects on strength and deformation of jointed rocks namely spacing, orientation, volume, friction, gouge mate-rial, ground water and internal pressure are considered in modifi ed joint factor (Jfg).

The factors, namely a and b, are readily estimated from core drill-ing data of the rock mass and geotechnical site investigation report. The variation of Jfg with RQD is shown in Figure 11.9. Zhang [41] and Trivedi [42] related rock mass modulus with RQD. We use the following relation [42] to fi nd deformation as described by:

Emr = exp [a Ch exp (b RQD)] (11.27)

11.2.5 Steps to Obtain Deformations from the Present Technique

The following steps should be followed to fi nd out deformation of cemented ash mass.

• Prepare bore log sheet containing important infor-mation, namely, depth of sampling at requisite inter-val, unit weight, overburden pressure, joint number

400

Lower bound

Lower Mean

Upper Mean

Upper bound

350

300

250

200

150

100

50

00 20 40 60

RQD

Jfg

80 100

Figure 11.9 Relationship of modifi ed joint factor and RQD [42].


appear in coring, orientation of joints, orientation parameter (nb), maximum, minimum and mean thick-ness of infi ll, density of infi ll, friction factor, water table, RQD, and columns for calculated inputs of Jfg, ap, Ch, Emr, and smr.

• Based upon inputs of Jfg and nb decide values of ap.• Evaluate the overburden and operate upon Ch

(Tables 11.9, 11.10; Fig.11.8).• Estimate values of Ch (as per Fig. 11.8 and Eqs. 11.19

and 11.20) and ap fi nd out Emr.• Draw a relationship between Jfg and RQD (Fig. 11.9) for

each data point to fi nd fi tting parameters a and b.• Using the Equation 11.25, Em can be found out to esti-

mate settlement of the footing embedded in rock mass as per the selected cases.

Table 11.9 Values of empirical parameters ‘ap’ and ‘Ch.’

Failure Mode

End Conditions

Limited Hardening

Applications

ap Ch

Rotation –0.025 –0.050 Slopes in cemented ash mass

Sliding –0.018 –0.036 Slopes in cemented ash mass

Splitting –0.0123 –0.025 Vertical cuts in cemented ash mass

Shearing –0.011 –0.022 Vertical cuts in cemented ash mass

Shearing –0.009 –0.018 Shallow depth fractured cemented ash mass

Shearing –0.008 –0.016 Shallow depth moderately fractured cemented ash mass

Shearing –0.005 –0.010 Shallow depth lowly frac-tured cemented ash mass

Shearing –0.003 –0.006 Fully cemented fractured cemented ash mass


Table 11.10 Value of hardening parameter for end conditions [42].

End conditions Coeffi cient (κ) and power (η) for the pressure ratio

ap κ η

–0.005 –0.040 –1.00

–0.009 –0.020 –1.25

–0.0100 –0.011 –1.42

–0.0125 –0.001 –2.00

11.3 Problems of Uncemented Ash Fill

11.3.1 Collapse, Piping and Erosion, Liquefaction

The uncemented ash fi lls are potentially vulnerable to excessive settlement [6, 47], collapse [48], fl ow failures [8], piping [49], ero-sion and liquefaction [10, 50]. Their potential vulnerability to fail-ure has a relation to their grain size parameter [48].

Terzaghi [49] described a piping failure as essentially induced by an excess head of water so that the fi ll adjoining the ash embank-ment remains in equilibrium, provided the hydraulic head hl is smaller than a certain critical value hc. However, as soon as this crit-ical value is approached the discharge increases more rapidly than the head, indicating an increase of the average permeability of the fi ll. Simultaneously the surface of the fi ll rises within a belt with a width of approximately n.D (n=0.5, as considered by Terzaghi [49] for sheet pile embedded to a depth of D) as depth of embedment D of the screen line of embankment. Finally a mixture of fi ll mate-rial and water breaks through the space located below the screen embankment. This phenomenon is called piping, and the hydraulic head at which piping takes place is the critical head. The piping beneath an ash fi ll is likely to cause a failure of the ash fi ll reten-tion system or ash pond embankment. The ash fi ll needs a factor of safety with respect to piping of the ash fi ll retention system after the water level has been lowered within the dam to a depth below the outside water level.

The process of piping is initiated by an expansion of the fi ll mate-rial between the buried portion of the impermeable ash embankment


and a distance of about n.D downstream from the screen line. This expansion is followed by an expulsion of the infi ll out of this zone. No such phenomenon occurs unless the water pressure overcomes the weight of the fi ll located within the zone of expulsion. It can be assumed that the body of fi ll, which is lifted by the water, has the shape of a block with a width n.D and a horizontal base at some depth m.D below the surface [m~1, for sheet pile embedded to a depth of D]. The rise of the block is resisted by the weight of the block and by the friction along the vertical sides of the block. At the instant of failure the effective horizontal pressure on the sides of the block and the corresponding frictional resistance are negligible. Therefore, the block rises as soon as the total water pressure on its base becomes equal to the sum of the weight of the block, fi ll mate-rial and water combined. The head hc at which the body is lifted is the critical head. The elevation of the base of the body is deter-mined by the condition that hc should be a minimum because pip-ing occurs as soon as the water is able to lift a block of fi ll regardless of where its base is located. The factors, namely “n and m,” depend upon friction, liquid limit, specifi c gravity and particle size similar to collapse. The critical hydraulic gradient (~0.3) for coeffi cient of uniformity is nearly equal to 10 for pond ash, making it potentially vulnerable to erosion.

Singh [50] analyzed the liquefaction characteristics of similar gradation of silty sand as that of ash fi lls [10]. The magnitude of lateral acceleration required for liquefaction of ash fi ll compared with sand (Table 11.11) in the same-sized horizontal vibration table (100 cm long, 60 cm wide, 60 cm deep) is signifi cantly lower than sand at a low (70%) relative density (0.31 to 0.14%), while at the higher (85%) relative density (0.51 to 0.31%). The increase in the liq-uefaction resistance of ash is reported [10] to be (1.84 to 1.70 times) higher compared to that in sand (1.10 times) for a similar increase in relative density (70 to 85%). In fact, the settlement of the ash fi ll was signifi cantly higher than sand.

The cyclic mobility-liquefaction characteristics of ash fi ll (obtained from the same source) in vibration table studies was evaluated by Trivedi et al. [10] and compared to clean sand. The ash fi ll settled nearly twice as much as sand for the equal vol-ume contained in the vibration chamber (1x0.6x0.6 m3) for nearly half of the dynamic disturbance. Further tests are recommended to verify predicted settlements of large-size footings on dynamic loads.


11.3.2 Collapse Behavior of Ash Fills

The general characteristics of collapsing fi lls are a sudden and a large volume decrease at a constant stress when inundated with water. According to Lutenegger and Saber [51] the collapse is asso-ciated with the meta-stable structure of a large open and porous fabric of the material. The earthen structures such as embankments, road fi lls and structural fi lls may collapse when the placement mois-ture content is dry of optimum. The infi ltration of the rainfall may be suffi cient to reduce the matric suction within the ash to a value low enough to trigger a shallow failure [52]. It has been observed that in a wide range of placement parameters the ash remains vul-nerable to the collapse on submergence in working stress range. A slip failure was reported at the ash dump of Vijayawada ther-mal power plant resulting in the destruction of several houses and the swamping of land with fl y ash. The sudden failure of a large fl y ash disposal dump after rainfall and the associated mudfl ow at Panki, Kanpur was reported [53]. Such failures are not quite repre-sentative of conventional failures. Several studies have indicated that compaction control of coal ash in the fi eld by usual methods is often poor. It adds to the vulnerability of ash fi ll to a wetting-induced collapse. The soils that exhibit collapse have an open type of structure with a high void ratio as expected in the case of ashes. According to Barden et al. [54] the collapse mechanism is controlled

Table 11.11 Range of parameters for liquefaction analysis [10].

Materials (Sand and Coal Ash)

Range Remarks

Relative density 65 to 85% By weight volume relation

Amplitude 1to15 mm Vibration table

Pore pressure at 5, 15, 25cm depth

At varied density and amplitude

At sinking of metal-lic coin

Settlement 5 to 20 mm At sinking of metal-lic coin

Number of cycles At varied density and amplitude

At sinking of metal-lic coin


by three factors: a potentially unstable structure, such as the fl occu-lent type associated with soils compacted dry of optimum or with less soils, secondly, a high applied pressure which further increases the instability, and a high suction which provides the structure with only temporary strength which dissipates upon wetting. As per an empirical study [1], the dry unit weight and water content are gen-erally considered as important parameters that control the collapse of metastable structure of soils, if the dry unit weight is less than 15 kNm3. The tentative dry unit weight of the coal ashes was often found to be less than 10 kNm3, suggesting a possibility of collapse.

Once a granular arrangement takes up a loose packing under a favorable condition of pressure and moisture the collapse of ash material occurs. This tendency is quantifi ed in terms of the contact separation parameter (D50/Da) defi ned in the Figures 11.10–11.12 and distance of placement void ratio to the minimum void ratio. Therefore, the collapse potential (Cp) and collapsibility factor (F) are:

Cp = Δh/h (11.28)

F= (ei - emin)/ emin (11.29)

where Δh is change in the sample thickness (h) upon inundation. Alternatively, ei is placement void ratio of granular materials and

1.000

0.100

Collapse susceptible soils, Cp > 0.01

Collapse susceptible ashes, Cp > 0.0075

Proposed limit for collapse potential of ashes in oedomerer test

Cp (corrected) at Dc = 80%

Cp (with soluble) at Dc = 80%

0.010

0.0010.001 0.010

Mean size (mm)

Co

llap

se p

ote

nti

al

0.100 1.000

Figure 11.10 Effect of grain sizes on collapse potential [8].


1.000 wc = 0%wc = 5%wc = 10%wc = 20%wc = 30%

Collapse susceptible soils, Cp > 0.01

Placement moisture content prior to wetting

Collapse susceptible ashes, Cp > 0.0075

0.100

0.010

0.0010 50 100

Stress level (kPa)

Co

llap

se p

ote

nti

al

150 200 250

Figure 11.11 Effect of moisture content on collapse potential [8].

D50 Da = 1 mm

Figure 11.12 Defi nition of contact separation parameter D50/Da [48].

emin is void ratio corresponding to maximum dry density in Proctor compaction. The variation of maximum and minimum void ratio of sand and ashes is empirically related with the grain sizes [15]. The void ratio extent defi ned by a difference of maximum and minimum void ratio drops with the increasing grain sizes [3]. This implies that the collapsibility increased by decreasing sizes. Therefore, the larger the value of F, the more the granular materials are predis-posed to collapse [8]. Figure 11.13 shows a reduction in collapsibil-ity factor F, with mean size in the loosest and the compacted states.

In the loosest state when grains are in progressive contact, a = 0.2 and b = 0.5.

F = a(D50/Da)–b (11.30)

On compaction the negative exponent “b” of grain separation parameter goes on reducing from 0.5 to one.


However, compared to the loosest state, all the granular materi-als reach nearly a common collapsibility level in a compacted state. At 90% degree of compaction, a collapsibility level is arrived at, which is associated with small volume change upon collapse that does not refl ect collapse. Moreover, it has the practical problem of precise measurement of the volume change. Thus, the variations in the measured collapse at 90% degree of compaction may forbid interpretation of any trend.

The collapsibility factor allows for assessment of the probable collapse. The collapse occurs if the sample attains a minimum void ratio on inundation. The maximum probable collapse potential is computed by:

Cpr = (ei-emin)/ (1+ei) (11.31)

where Cpr is the maximum probable collapse potential, ei is void ratio in a loose state and emin is void ratio corresponding to a maxi-mum dry density in Proctor compaction. The collapse potential shows that the decreasing mean size tends to reduce the difference between the maximum probable and the observed collapse at 80% degree of compaction. While at 90% degree of compaction a signifi -cant scatter of the data is observed.

3.00F in loosest stateF at Dc = 80%F at Dc = 90%Anticipated F2.50

2.00

1.50

1.00

0.50

0.000.001 0.010 0.100

D50/Da

Co

llap

sib

ility

fac

tor

F

1.000

Figure 11.13 Criteria of collapse based on collapsibility factor [8].


Because of the above observations the classifi cation of granular materials at 80% degree of compaction was found to be appropri-ate for the evaluation of collapse. The mean particle size was seen to control the collapse of granular materials. If the mean size was greater than 1mm the granular materials were non-collapsible and others were collapsible under specifi c conditions. The value of col-lapse potential in the critical range of stress and moisture was 3 to 6 times that of the corresponding dry condition (Figure 11.11). This sug-gested susceptibility of a non-collapsible dry granular material to the collapse in partly saturated condition. In order to obtain the value of collapse potential of partly wet granular materials, a multiplier may be applied. The vulnerability of fi ne ash fi ll is validated by conduct-ing triaxial tests on partly saturated ash compacted at maximum dry density and saturated ash. The deformation modulus of saturated ash is signifi cantly lower than the partly saturated ash (Figure 11.14).

The collapsible granular materials were further divided into the granular materials of low, medium and high collapsibility based on their collapse potential. The collapsible and the non-collapsible granular materials were identifi ed using a model collapse test on selected samples. Normally, the weight of a particle of a natural soil

20000

Ept

Esat

Linear (Ept)

Linear (Esat)Power (Ept)

Power (Esat)

15000

10000

5000

00 50 100

Effective confining pressure (kPa)

Def

orm

atio

n m

od

ulu

s (k

Pa)

150 200

Figure 11.14 Deformation modulus of fi ne saturated and partly saturated ash based on triaxial test.


of similar grain size is 1.5 to 1.3 times that of ash materials. These soils remain stable at or less than 1% volume change (Cp = 0.01). Being light in weight, the ash material has a propensity to be unsta-ble in the presence of buoyancy, which plays a role in the model and the fi eld conditions. Therefore, among the lightweight granu-lar materials particles, 0.75% volume change (Cp = 0.0075) triggered collapse failure in the fi eld. Coincidentally, 1% volume change of soils is 1.3 times that of the limit recognized for the collapsible ash materials.

It was observed that the sand and the coarse ash had very close value of the median size. It being a granular material collected dry, having around 25% particles in silt range, it had a higher collapse potential than the sand. It was recognized that all the collapsible granular materials had relatively more fi nes. Among coarse-grained granular materials, a scatter in collapse potential was observed. A relationship between corrected collapse potential at 200 kPa (Dc = 80%) and mean particle size is obtained with a satisfactory coeffi cient of determination. The collapse potential is expressed by:

Cp = n (D50/Da)m (11.32)

where Cp is collapse potential of a granular material, D50 is mean particle size in mm, and Da is reference size = 1 mm, and m and n are fi tting constants for the granular materials.

11.4 Ash as a Structural Fill

There is only scant data available on the interpretation of the load-bearing behavior of ash fi lls. The penetration test results ana-lyzed by Cousens and Stewart [12, 57] and Trivedi and Singh [17] showed the scope of development of new correlations for evalua-tion of foundation settlements on coal ash. Leonards and Bailey [30] favored the use of plate load test results for coal ash. Trivedi and Sud [20, 22] examined the evaluation of bearing capacity and settle-ment of ash fi lls. The present work reviews the plate settlement on coal ash for working out a strategy for evaluation of foundation settlement. Some of the case studies are reported on the investiga-tion and assessment of load-bearing behavior of coal ash, which is also relevant to its uses in embankment fi lls. These tests are namely penetration and plate load tests as described below.


11.4.1 Penetration Test

The standard and cone penetration devices are used to evaluate the stability of a landfi ll. The standard penetration test is used in dif-ferent parts of the world with a slight variation in the version of its use. It involves an estimation penetration value of 300 mm run of a split barrel of 50(±2) mm external and 35(±1) mm internal diameter under impact of 63.5 kg hammer The results of standard penetration tests (SPTs) conducted on hydraulically deposited Illinois ash were reported. The penetration was observed to be of several stretches of 30 mm under the weight of drill rods to 9 blows per 300 mm of penetra-tion. The SPT value for compacted Kanawha ash was observed among 10 and 31 and had an average of 19.5 for seventeen independent tests conducted in four borings, excluding the values obtained for lenses of bottom ash. Dry density values determined for fi ve of the nine Shelby tube samples taken from these bore holes suggest hardly any correla-tion with the N-values (fi eld density 95% to 100% of maximum dry density ~14 to 16 kN/m3). The average cone penetration resistance is related to SPT value (= 200 x N kNm2), where N is the SPT value. The ashes normally in silt size, especially in loose condition, may liquefy under the tip of penetrometer below the water table resulting in a lower penetration record. For the compacted ash fi lls, the density pro-jection and SPT (N) number are presented in Table 11.12. These low

Table 11.12 Variation of SPT value for ash deposits.

Ash Type Dc (%) N-value Reported by

Compacted Kanawaha ash 95–100 10–31 [9]

Hydraulically deposited Illinois ash

Loose state Zero [9]

Well compacted Ontario fl yash 85–100 10–55 [56]

Bottom ash – 75 [56]

Compacted ash dyke 95 4–27 [24]

Hydraulically deposited ash Loose state Zero [24]

Well-compacted fl y ash in a valley-Pittsburg

– 75 [57]

Hydraulically deposited ash Loose state Zero-1 [1]

Hydraulically compacted ash Dense state 1–10 [1]


values of N are partly associated with a low unit weight and partly with a high percentage of fi nes and a high moisture condition. A high value of N is associated with the presence of lenses of cemented ash and partly with a higher compaction and aging.

An average friction ratio for sleeve and ash is from 3 to 5 times higher than the value cited by Schmertmann [55] for clay silt sand mixes, silty sands, silts and sands. Toth et al. [56] quoted a wide variation in SPT value ranging from 10–55 in fl y ash with angle of internal friction ranging from 35 to 36°. The empirical co-relation between SPT values and φ (angle of internal friction) for natural soils in this range of SPT value (N = 10 to 55) indicated angle of internal friction ranging from 30 to 45°. The investigations carried out by Cousens and Stewart [12] for the range of cone resistance and the friction ratio (200 kPa and 8%, respectively) indicated grain sizes in the range of silt (60–80%) and clay (5–10%). For a tar-get relative density (50 to 85%), variation in standard cone resis-tance ranges from 2000 to 6000 kPa. However, the average friction (between sleeve/cone and ash material) ratio was observed to be 3 to 5% [15, 9]. The settlement of these ash fi lls on the basis of the Schmertmann method was found to be a non-conservative estimate.

11.4.2 Load Test

On the basis of a case study on Indianapolis ash, Leonards and Bailey [30] suggest that the load-settlement relation for the founda-tion on compacted ash cannot be inferred from standard penetra-tion or static cone penetration tests. This is largely attributed to the inadequacy of penetration tests to sense the effect of compaction related to pre-stressing of the coal ash. The predicted settlements for a selected footing 2.1-m wide at design pressure of 239 kPa on well-compacted ash from the data of standard penetration (SPT), cone penetration (CPT) and plate load (PLT) tests are presented in Table 11.13.

Table 11.13 Predicted settlement for well-compacted area [30].

Testing Technique CPT SPT PLT

Settlement (mm) 15 25.4 5.08


blows/305 mm penetration) indicated that ash materials are sig-nifi cantly less compressible in the pressure range of interest. At 100 kPa, compacted ash may settle less (0.6 mm) compared to the settle-ment of the same plate on sand (approximately 1.2 mm).

Toth et al. [56] reported a case study on the performance of Ontario ash (a typical class F ash). During the compaction of ash for landfi ll, it was observed that the densities being achieved in the fi eld were normally below 95% of maximum Proctor density. On the basis of the good bearing capacity observed in the plate load test, 95% of maximum Proctor density is recommended as the tar-get density for the fl y ash landfi lls. Toth et al. [56] obtained short- and long-term test results for circular plates of 0.3-m and 0.6-m in diameter. The settlements for long-term tests occurred within the fi rst hour of load application.

The results of investigations by various investigators are given in Table 11.14. The tests conducted on a common degree of com-paction (93.4%) and plate size (600-mm diameter), had different settlement records at a stress level of 100 kPa (2.61 and 3.59 mm). Trivedi and Sud [3] have shown that the variation in grain sizes

Table 11.14 Settlement of test plate on compacted ash fi ll.

Plate size (mm) and shape

Degree ofCompaction

(%)

Moisture Content

(S/B)%at 100 kPa

Interpolation Data

900, square 85.24 Wet of Critical

0.56 [22]


0.63 [22]


0.45 [22]


0.35 [22]

300, square 85.24 Dry of Critical

1.03 [22]

300, square 81.55 Dry of Critical

1.56 [22]


of ashes may result in different settlement characteristics even at a common degree of compaction. Trivedi and Singh [9] reported a higher load-bearing capacity of ash fi lls than actually estimated by cone resistance.

11.4.3 Test Setup for Ash Fills and Testing Technique

The ash is normally deposited in loose lift of 150 mm in a trench of plan dimension of 1.5 m × 1.5 m. It is compacted by a pre-calibrated plate vibrator mounted on a fl at rectangular plate (152 mm x 390 mm). The rating of the plate vibrator is kept at 3000 rpm. A constant magnitude of vibration is required to achieve the desired relative

Plate size (mm) and shape

Degree ofCompaction

(%)

Moisture Content

(S/B)%at 100 kPa

Interpolation Data


0.40 [22]


0.34 [22]

600, circular 93.40 Wet of Critical

0.430.50

[56]


0.23 [56]


0.15 [56]

300, square,Long term

– Wet of Critical

0.15 [56]

600, square < 95% Wet of Critical

0.22 [30]

300, square < 95% Wet of Critical

0.23 [30]

300, square Sand, N = 50 – 0.36 [30]

Table 11.14 (Cont.)


density. The trench is fi lled up in layers maintaining constant den-sity throughout. The density checks are applied at regular intervals using thin core cutter sampling and penetration of an 11-mm diam-eter needle penetrometer under constant pressure .

The plate load test is conducted on the compacted ash fi ll. A few model tests (Table 11.15) are carried out on surface footings (0.1 to 1 m wide plates) in dry as well as submerged conditions for dif-ferent ashes and a sand to check the reproducibility of the results. Additionally, on site density checks and laboratory shear tests are also carried out. The displacement of the plate is monitored using pre-calibrated settlement gauges of least count 0.01 mm. The total assembly including hydraulic jack, proving ring and the plate are aligned with the help of a plumb bob to attain verticality.

The load capacity of ash fi ll is estimated by conducting load tests using different plates on ashes at varying degree of compaction. A summary experimental program is considered. An average of at least two tests is considered to reach a common load settlement plot

Table 11.15 Summary of load tests used for prediction.

Ash type

Test conditions

Size (m)

Shape Dc(%) No. of tests

Max. Pressure

Sand Dry 0.1 Strip – 2 Failure

A1 Dry of Critical

0.1 Strip – 2 Failure

Wet of Critical

0.1 Strip 2 Failure

A2 Dry of Critical

0.1 Strip – 2 Failure

Wet of Critical

0.1 Strip 2 Failure

A2 Dry of Critical

0.30.6

Square 81.5585.24

44

Failure Failure

Wet of Critical

0.30.60.9

Square 85.2490.29

444

200 kPa160 kPa100 kPa


if the values are within the range of 10%. The results evaluated from typical pressure settlement plots are shown in Figures 11.15–11.17.

A plate of the desired size is placed on the ash fi ll. A leveled 10-mm thick layer of dry ash is spread on compacted ash to ensure relatively complete and uniform contact between the bearing plate and compacted ash. The plate is loaded with the hydraulic jack against a reaction truss. After application of seating load, the load is increased in regular increments. Bearing plate settlement is mea-sured with an accuracy of 0.01 mm.

A1–0.1 Dry

A1–0.125 Dry

A1–0.1 Submerged

A2–0.1 Submerged

A2–0.3 Dry

A2–0.1 Dry

A2–0.125 Dry

A2–0.1,D/B = 1

A2–0.125,D/B = 1

400

600N

γγ

200

00

Ir

1 2 3 4 5 6

Figure 11.15 Bearing capacity factor for coal ashes based on relative dilatancy [20].

Increasing settlement ratio

Increasing size

1

0.5

00 1 2 3 4 5

Increasing depth

Ir

I pr

I- At settlement ratio of 20%

II- By double tangentIII- Ir (Perkin and Madson, 2000)

Figure 11.16 Progressive failure index and relative dilatancy for ashes [20].


Each load increment is maintained on the bearing plate as long as no change in the settlement is observed for two hours in succes-sion. Maximums of load required for failure of the deposit were estimated by the bearing capacity factors obtained from small-scale tests. A sharp increase of bearing plate settlement is considered as an indication of the beginning of the ash failure phase. The settle-ment observation under the fi nal load is taken to the maximum of 24 hours.

11.4.4 Bearing Capacity of Ash Fill

The generally agreed upon bearing capacity equation for shal-low depths uses bearing capacity factors Nc and Nq. However, substantial differences have been reported in the semi-empirical bearing capacity factor for shallow foundations Nγ in numerous studies [9, 20, 59].

The bearing capacity equation for smooth strip foundations on the surface of a fi ll with surcharge, is given by:

qult = C Nc + σov’ Nq + 0.5 Nγ γ ’ B (11.33)

where σov’ is the effective surcharge pressure acting at footing base expressed in terms of effective stress, γ’ is buoyant unit weight and B is footing width; Nc, Nq and Nγ are the bearing capacity factors.

Pressure (kPa)

00 50 100 150 200 250

5

10

Set

tlem

ent

(mm

)

Figure 11.17 Pressure settlement plot interpreted for ashes from varied sources.


The value of Nq is obtained using a simple plasticity theory for a weightless soil;

Nq = tan2 (π/4 + φ’/2) eπtanφ’ (11.34)

The bearing capacity obtained from the Equation 11.33 does not increase linearly with the width of footing or overburden. This phe-nomenon was frequently termed as scale effect by De Beer [59], who attributed it to the nonlinear shape of soil failure envelope result-ing in secant measure of friction angle, which decreases with mean normal effective stresses. With increasing confi nement, dense and loose cohesionless soils have much less of a marked difference in peak angle of internal friction. This effect is pronounced in the geo-materials, namely ash fi lls, which suffer from progressive crushing.

Experimentally, Nγ for a surface footing without surcharge is obtained as:

Nγ = qult/ (0.5 B γ Sγ) (11.35)

where Sγ = shape factor, which is not required in the relative dilat-ancy approach, and B = width of the strip footing, and γ is the unit weight of the ash. Since φ varies as the state of stress, density and material characteristics of the soil, the concept of stress dilatancy developed by Bolton [27] is utilized. He proposed the empirical equation:

fpeak =fcr + A Ir (11.36)

Ir = RD (Q – ln p¢) – r (11.37)

where A is an empirical constant and has a value of 3 for triaxial case; Ir is relative dilatancy index; p¢ is effective mean confi ning pressure in kPa; RD is relative density; Q and r are empirical mate-rial fi tting constants with a value of 10 and 1, respectively, for clean silica sand. Incorporating the triaxial test data, Bolton [27] sug-gested that progressive crushing suppresses dilatancy in the soils of weaker grains, i.e., limestone, anthracite, and chalk, which show “Q” values of 8, 7, and 5.5, respectively. The ash containing sub-stantial amount of crystalline fi ne silica grains followed by alumina and oxides of iron, calcium and magnesium may have Q as low as 7.7 [20]. This occurs mainly because of the reduction of critical


mean confi ning pressure beyond which increase in mean confi ning pressure for a relative density does not increase peak angle above critical angle. Perkins and Madson [59] proposed to integrate this approach of progressive failure with the bearing capacity of shal-low foundations on sand. This approach is extended to meet the requirements of the plate load test on an ash fi ll.

The ultimate bearing capacity of the surface footing on ash fi lls is proposed to be estimated by:

qult = 0.5 Nγ g’ B (11.38)

where B is the width of the footing and g’ is effective unit weight of the footing.

The value of bearing capacity factor Ng (Fig. 11.15) may be esti-mated [20] by:

Ng = 20e ζ Ir (11.39)

The typical value of ζ = 0.5 to 1, which depends upon progressive failure Ipr (Fig. 11.16) obtained from relative dilatancy, and fpeak and fcr are the peak and constant volume angle of internal friction. A=5 and 3 for plain strain and triaxial conditions respectively.

p - ( )/A ln( / ) cQ RD r RD p paj j= − +′ ′ (11.40)

where Q = 10 for sand and 7.7 for coal ash. Rd is relative density and p’ is effective mean confi ning pressure below the footing.

The ultimate bearing capacity of ash fi lls was observed among the estimates of critical and peak friction angles. The bearing capac-ity of ash fi ll evaluated is always lower than that obtained by the use of peak friction angle. Figure 11.16 shows the variation in the index of progressive failure (Ipr), with relative dilatancy index for a surface and an embedded footing. The index of progressive failure is hereby defi ned as:

Ipr = [qult (at f’ peak) – qult (from cone penetration test)] / [qult (at f’ peak) – qult (at f critical)] (11.41)

If Ipr takes a value of 1, it implies that ultimate bearing capacity of ash fi ll is governed by critical friction angle, while a value of zero


indicates the peak angle of friction is fully mobilized. The concur-rence of a relatively high value of factor (Q – ln p ) at peak cone resistance in ash fi lls leads to higher values of relative dilatancy index among ashes.

11.4.5 Settlement of Ash Fills by PLT

Trivedi and Sud [22] presented an evaluation of the settlement of ash fi lls. In order to investigate settlement characteristics of compacted coal ash, fi eld plate load tests were analyzed for ash compacted to varying degrees of compaction and plate sizes (Table 11.16).

It is observed that ash may be compacted to the same degree of compaction at two moisture contents, one dry of critical and the other wet of critical moisture content. The critical moisture content is defi ned as the moisture content or range of moisture content in which vibratory effort becomes ineffective and ash bounces back to a loosest packing corresponding to which dry unit weight of ash is minimum in the presence of moisture.

On the dry side of critical, ash packing is very sensitive to mois-ture. Within the limitation of workability in fi eld, different degrees of compaction were selected (i.e., 85 and 80%). The observations of the moisture density relationship in a fi eld are similar to that in a laboratory vibration test. The increasing moisture content from 5 to 10% decreases the degree of compaction from 85 to 80%. Further, the settlement of 300-mm x 300-mm test plate increases from 3 to 5 mm at 100 kPa.

The coal ash is compacted to a higher degree of compaction (90%). By increasing degree of compaction from 85 to 90% on the wet side of critical, settlement was reduced from 1.5 to 1 mm. That

Table 11.16 Effect of plate size on settlement at wet side of critical [47].

Plate Size (mm) Settlement at 100 kPa in mm

Dc = 90% Dc = 85%

300 × 300 1.05 1.45

600 × 600 2.4 3.8

900 × 900 3.1 5.1


is an improvement in the degree of compaction by 5% (from 85 to 90%), and the settlement is reduced by one-third.

11.4.6 Settlement on Ash Fills by PLT, CPT and SPT

The case studies [30, 56, 20, 22] have shown that standard penetra-tion test results might overestimate settlements of ash fi ll as high as fi ve times that of predicted value by plate load test, while the cone penetration test overestimated settlements as high as three times [30]. The plate load tests tend to give a more precise indication of actual settlements of larger sizes. The observed data of several investigators [30, 56] is given along with the results of the investiga-tions in Table 11.17.

From the analysis of data of the present investigation and that published by Leonards and Bailey [30], it is understood that well-compacted ash for the footing size and stress level of interest, has settlement directly proportional to pressure up to 200 kPa. Therefore the settlements at 100 kPa are interpolated from the data published [30, 56, 20, 22].

The critical moisture content is defi ned as moisture content at which ash attains minimum density when compacted by vibra-tion. The moisture-density curve is almost symmetrical about this moisture content. Therefore, ash may be compacted at two different moisture contents: one dry of critical and the other wet of critical.

Table 11.17 Predicted settlements (mm) using Terzaghi and Peck for-mula [22].

Bf (m) Dc ~ 85% [22]

Dc ~90% [22]

Dc < 95%, [30]

Dc ~ 98.2%, [56]

0.60 2.56 1.85 1.24 0.79

0.90 3.26 2.36 1.575 1.07

1.2 3.71 2.68 1.79 1.15

1.5 4.01 2.90 1.94 1.24

1.8 4.25 3.07 2.05 1.31

2.1 4.43 3.21 2.14 1.37


Adding moisture beyond this critical moisture content, appar-ent cohesion develops which impedes the deformability of ash. This apparent cohesion is destroyed gradually by addition of water beyond optimum moisture. It seems that owing to the development of apparent cohesion, ash becomes far less deformable above a con-stant dry density (Dc = 85%).

The settlement record at 100 kPa suggest that, on the dry side of critical, settlement is more than two times that of the wet side of critical. The percentage increase in settlement by compacting ash at dry side of critical, instead of wet side of critical, at 100 kPa on 0.3m x 0.3m square plate at Dc of 85% is 115%.

There is a signifi cant impact of degree of compaction on the dry side of the critical. For a drop in the degree of compaction from 85 to 80% (decrease in degree of compaction is 4.32%) there is an increase of settlement at 100 kPa from 3.1 to 4.7 mm (increase in settlement is 51.61%). Similarly on the wet side of critical, for a decrease in the degree of compaction from 90 to 85% (percent-age decrease in degree of compaction is 5.6%) there is an increase in settlement from 1.05 to 1.45 mm (percentage increase in settle-ment is 38%).

The scale effects are clearly visible in the settlement of ash deposits. Plots are drawn from experimentally observed and pre-dicted settlement for 0.3m × 0.3m, 0.6m × 0.6m and 0.9m × 0.9m square size plates, respectively. Using settlement of 0.3m × 0.3m plate as plate settlement, settlement of footing is estimated by the formula [58]:

( )( )

20.3

0.3

f p

f p

p f

B Bs S

B B

⎡ ⎤+⎢ ⎥=⎢ ⎥+⎣ ⎦

(11.42)

where Bp = width of plate in meter; Bf = width of footing in meter; Sp = settlement of plate in mm; and Sf = settlement of footing in mm.

The predicted settlements underestimated actual settlement at all degrees of compaction, even at a high degree of compaction (98% and at less than 95%). The settlement of 0.6 m (least dimen-sion) plate has been underestimated by 45% and 7.5%, respectively. In the investigation [47] the predicted settlement from the actual settlement of 300-mm plate underestimated the actual settlement of


0.6 m plate (least dimension) at 90 and 85% degree of compaction by 56 and 62%, respectively.

The settlement of ash fi ll at varying relative densities is obtained by CPT (Figure 11.18) using the Meyerhof [1] method modifi ed by Trivedi and Singh [9]. The following relation gives the settlements of ash fi ll which is less than fi fty to less than ten percent of the settlement obtained by the Meyerhof [1] method.

Sc= ΔpB/[n*RD + m)]qc (11.43)

where Δp = net foundation pressure. The cone resistance (qc) is taken as the average over a depth equal to the width of the footing (B) and RD is relative density, while n and m are constants which take a value ~ 10 to 3.5, respectively.

11.4.7 Settlement of Footings on Ash Deposit

The predicted settlements according to the Terzaghi and Peck [58] extrapolation is not in agreement of settlement of footings larger than 0.6 m (least dimension) on compacted ash fi ll (Figure 11.19). The predicted settlements based on actual settlement of 300-mm square plate seriously underestimate the observed settlements using method by D’Appolonia et al. [60].

00

400

800

1200

Dep

th (

mm

)

5 10 15qc (MPa)

20 25

Sand; RD = 50%Sand; RD = 65%Sand; RD = 85%Ash; RD = 51.6%Ash; RD = 77.4%Ash; RD = 85.5%

Figure 11.18 Cone resistance plot interpreted for ash and sand.


Table 11.18 Underestimation of settlement by the Terzaghi and Peck formula [47].

Dc (%) % Under estimation of settlement

Interpreted from the reference

98.2 44.36 [56]

< 95 7.46 [30]

90.29 56.25 [47]

85.24 61.84 [47]

14

Dc = 98%

Dc = 90%

Dc = 85%

12

10

8

6

4

2

00 1 2 3

Footing width (m)

Set

tlem

ent

(mm

)

4 5 6

Figure 11.19 Predicted settlements as per Terzaghi and Peck criteria [22].

The mean value of ratio of predicted settlement according to the Terzaghi and Peck extrapolation and experimentally observed settlements was found to be 0.3. Table 1.18 presents percentage of underestimation of (0.6m x 0.6m) footing settlement by the Terzaghi and Peck formula at varying degrees of compaction.

The predicted settlements according to the criterion suggested by D’Appolonia et al. [60] is estimated at 100 and 200 kPa. The experimental data for varying sizes of footing at probable degree


of compaction is plotted in Figures 11.19 and 11.20. The expected settlement, at a pressure of 100 kPa, indicates least possibility of exceeding the allowable limit of settlements in the probable degree of compaction.

The settlement estimate corresponding to the PLT values obtained for the ash fi lls is estimated for 1m wide footing as per Terzaghi and Peck method as shown in Figures 11.19 and 11.20, which shows excessive rigidity of ash fi lls to settlement in partly saturated con-dition. The settlement estimate corresponding to the SPT values obtained for the ash fi lls is estimated for 1m wide footing at 100 kPa as per Meyerhof [61] and Burland and Burbage [62] methods (shown as Mh and B&B, respectively, in Figure 11.21), which may not show excessive vulnerability of ash fi lls to settlement in satu-rated condition. The settlement estimate corresponding to the CPT values obtained for the ash fi lls is estimated for 1m wide footing at 100 kPa (Figure 11.22) which shows comparison of ash fi lls to settle-ment with PLT (Table 11.19).

10

Dc > 95%

Dc > 98%

Dc > 90%

Dc > 85%

9

8

7

6

5

4

3

2

1

00.1 1

Footing Size (m)

Set

tlem

ent

(m)

10

Figure 11.20 Predicted and observed settlement at varying degrees of compaction and footing width [22].


160

Saturated settlement (SM-B&B)

Saturated settlement (SM-Mh)

Settlement (SM-B&B) at 100 kPa B = 1

Settlement (SP-Mh)

Settlement (SM-Mh)

120

80

40

00 2 4 6

N60

Set

tlem

ent

(mm

)

8 10

Figure 11.21 Predicted settlements as per SPT.

0.10

2

Pre

dic

ted

set

tlem

ent

(mm

) u

sin

g d

efo

rmat

ion

mo

du

lus

fro

m C

PT

Dc = 85.6%Dc = 90.2%Dc = 92.4%Dc = 95.5%

4

6

8

10

1 10Footing width (m)

Figure 11.22 Predicted settlement from CPT [9].


11.5 Conclusions

The analysis in this chapter comprehensively provides a design focused upon the engineering behavior of ash fi ll so that it can ful-fi ll its intended function from a structural and utilization point of view. The applicability of RQD and its correlation with strength parameters of geomaterials is available in the literature. The use of RQD techniques for the evaluation of cemented ash having refusal to penetrate standard cone and SPT may provide a useful insight into the behavior of cemented fi ll. The bearing capacity and settle-ment of footing on weakly compacted ash is relatively critical com-pared to that degree of compaction at the wet side. The extent of progressive failure below the footings at varied degrees of compac-tion can be estimated from the relative dilatancy considerations.

The settlement of fi ll is worked out as per load tests, SPT and CPT at the favorable degree of compaction, for intended footing size and at desired stress level for the ash compacted on the wet side of the critical. The pressure corresponding to safe settlement may also be ascertained from the available data.

From the estimates of safe loads and safe settlements, allowable bearing pressure may be estimated.

The submergence of fi ne ash deposit is critical to its stability against piping, collapse and liquefaction, hence, effects of water on ash fi ll (fi ner than 75μ) need to be critically examined for loose uncemented ash fi lls.

Salutations, Acknowledgement and Disclaimer

The present chapter is compiled from the work of numerous investigators on varied geomaterials. The conclusions drawn from different resources are interpretive in nature and collectively

Table 11.19 Predicted settlements according to modifi ed criteria [47].

Degree of Compaction Dc (%)

Settlement (in mm) at 100 kPa 3m wide footing

Interpreted from the reference

85.24 9.3 [47]

90.29 5.6 [47]

98.20 3.7 [56]


provide a useful insight into the engineering behavior of ash fi ll. This framework is drawn from laboratory and fi eld data that are predominantly available in published form; however, fur-ther endorsement through a comprehensive testing program is recommended.

References

1. A. Trivedi, Engineering behavior of coal ash, Ph.D. Thesis, Dept. Civil Eng., TIET, Patiala, 1999.

2. http://en.wikipedia.org/wiki/Fly_ash, 2013. 3. A. Trivedi and V.K. Sud, Grain characteristics and engineering proper-

ties of coal ash, Granular Matter, Vol. 4, No. 3, pp. 93–101, 2002. 4. Federal Register, Vol. 65, No. 99, pp. 32214–32237, 1989. 5. K. Prakash and A. Sridharan, Benefi cial properties of coal ashes and

effective solid waste management, Practice Periodical of Hazardous, Toxic, and Radioactive Waste Management, Vol. 13, No. 4, pp. 239–248, 2009.

6. A. Trivedi and V.K. Sud, Effect of grain size on engineering properties of coal ash, Indian Journal of Environmental Protection, Vol. 23, No. 5, pp. 525–553.

7. M.G. Ayoola and V.O. Ogunro, Leachability of compacted aged and fresh coal combustion fl y ash under hydraulic fl ow conditions, in: GeoCongress 2008, American Society of Civil Engineers, 2008, pp. 692–699.

8. A.Trivedi and V.K.Sud, Collapse behavior of coal ash, J. Geotech. Geoenviron. Eng., Vol. 130, No. 4, pp. 403–415, 2004.

9. A. Trivedi and S. Singh, Cone resistance of compacted ash fi ll, Journal of Testing and Evaluation, Vol. 32, No. 6, pp. 429–437, 2004.

10. A. Trivedi V.K. Sud and R. Pathak, Liquefaction characteristics of coal ash, in: Fly Ash Characterization and Its Geotechnical Applications, 1999, pp. 189–194.

11. A. Dey and S.R. Gandhi, Evaluation of liquefaction potential of pond ash, in: Geotechnical Engineering for Disaster Mitigation and Rehabilitation, H. Liu, A. Deng, and J. Chu, Eds. Springer Berlin Heidelberg, 2008, pp. 315–320.

12. T. Cousens and D. Stewart, Behaviour of a trial embankment on hydraulically placed pfa, Engineering Geology, Vol. 70, No. 3–4, pp. 293–303, 2003.

13. N.A. Lacour, Engineering characteristics of coal combustion residu-als and a reconstitution technique for triaxial samples, Virginia Polytechnic Institute and State University, 2012.

14. S. Sharma, Ash dam damaged, NTPC’s production from plant hit, Economic Times, Sept 27, 2011.


15. A. Trivedi, S. Singh, and C. Singh, Characterization and penetration resistance of ash fi ll, Indian Journal of Environmental Proctection, Vol. 23, pp. 768–773, 2003.

16. A. Sridharan, N.S. Pandian, and S. Srinivas, Compaction behaviour of Indian coal ashes, Ground Improvement, Vol. 5, No. 1, pp. 13–22, 2001.

17. A. Trivedi and S. Singh, Geotechnical and geoenvironmental prop-erties of power plant ash, Journal of the Institution of Engineers. India. Civil, Vol. 85, pp. 93–99, Aug. 2004.

18. A.K. Choudhary, J.N. Jha, and K.S. Gill, Laboratory investigation of bearing capacity behaviour of strip footing on reinforced fl yash slope, Geotextiles and Geomembranes, Vol. 28, No. 4, pp. 393–402, 2010.

19. D.P. Mishra and S.K. Das, A study of physico-chemical and mineralogi-cal properties of Talcher coal fl y ash for stowing in underground coal mines, Materials Characterization, Vol. 61, No. 11, pp. 1252–1259, 2010.

20. A. Trivedi and V.K. Sud, Ultimate bearing capacity of footings on coal ash, Granular Matter, Vol. 7, No. 4, pp. 203–212, 2005.

21. R. Gupta and A. Trivedi, Effects of non-plastic fi nes on the behavior of loose sand-An experimental study, Electronic Journal of Geotechnical Engineering, Vol. 14, pp. 1–15, 2009.

22. A. Trivedi and V.K. Sud, Settlement of compacted ash fi lls, Geotechnical and Geological Engineering, Vol. 25, No. 2, pp. 163–176, 2007.

23. P.K. Mehta and P.J.M. Monteiro, Concrete: Microstructure, Properties, and Materials. McGraw-Hill, 2006.

24. U. Dayal and R. Sinha, Geo Environmental Design Practice in Fly Ash Disposal & Utilization. Allied Publishers, 2005.

25. S. Ojha and A. Trivedi, Shear strength parameters for silty-sand using relative compaction, Electronic Journal of Geotechnical Engineering, Vol. 18, No. Bund A, pp. 81–99, 2013.

26. R. Salgado, P. Bandini, and A. Karim, Shear strength and stiffness of silty sand, J. Geotech. Geoenviron. Eng., Vol. 126, No. 5, pp. 451–462, 2000.

27. M. Bolton, The strength and dilatancy of sands, Geotechnique, Vol. 36, No. 1, pp. 65–78, 1986.

28. S. Ojha, P. Goyal, and A. Trivedi, Non-linear behaviour of silty sand from catchment area of Yamuna River, in: Indian Geotechnical Conference, 2012, pp. 277–280.

29. S.K. Rao, L.K. Moulton, and R.K. Seals, Settlement of refuse landfi lls, in: Proceedings of the Conference on Geotechnical Practice for Disposal of Solid Waste Materials, University of Michigan, Ann Arbor, Michigan: June 13–15, 1977.

30. G.A. Leonards and B. Bailey, Pulverized coal ash as structural fi ll, J. Geotech. Eng. Div., ASCE, Vol. 108, 1982.

31. A.R. Cooper Jr and L.E. Eaton, Compaction behavior of several ceramic powders, Journal of the American Ceramic Society, Vol. 45, No. 3, pp. 97–101, 1962.


32. J.W. Hilf, Compacted fi ll, in: Foundation Engineering Handbook, Springer, 1991, pp. 249–316.

33. Y.S. Chae and J.L. Snyder, Vibratory compaction of fl y ash, in: Geotechnical Practice for Disposal of Solid Waste Materials, 1977, pp. 41–62.

34. S.K. Saxena and R.M. Lastrico, Static properties of lightly cemented sand, Journal of the Geotechnical Engineering Division, Vol. 104, No. 12, pp. 1449–1464, 1978.

35. A. Puppala, Y. Acar, and M. Tumay, Cone penetration in very weakly cemented sand, Journal of Geotechnical Engineering, Vol. 121, No. 8, pp. 589–600, Aug. 1995.

36. F. Schnaid, P. Prietto, and N. Consoli, Characterization of cemented sand in triaxial compression, J. Geotech. Geoenviron. Eng., Vol. 127, No. 10, pp. 857–868, Oct. 2001.

37. N. Consoli, A. Viana da Fonseca, R. Cruz, and K. Heineck, Fundamental parameters for the stiffness and strength control of artifi cially cemented sand, J. Geotech. Geoenviron. Eng., Vol. 135, No. 9, pp. 1347–1353, Feb. 2009.

38. A. Trivedi and N. Kumar, Foundation settlement for footings embed-ded in rock masses, in: UKIERI Concrete Congress - Innovations in Concrete Construction, 2013, pp. 2031–2039.

39. A. Trivedi and N. Kumar, Strength of jointed rocks with granular fi ll, in: ISRM International Symposium - 6th Asian Rock Mechanics Symposium, October 23 - 27, 2010, New Delhi, India, 2010.

40. Z. Sen, RQD models and fracture spacing, Journal of Geotechnical Engineering, 1984.

41. L. Zhang, Method for estimating the deformability of heavily jointed rock masses, J. Geotech. Geoenviron. Eng., Vol. 136, p. 1242, 2010.

42. A. Trivedi, Estimating In Situ Deformation of Rock Masses Using a Hardening Parameter and RQD, Int. J. Geomech., ASCE, 13(4), 348–364, 2013.

43. T. Ramamurthy and V. Arora, Strength predictions for jointed rocks in confi ned and unconfi ned states, International Journal of Rock Mechanics and Mining Sciences, Vol. 31, No. 1, pp. 9–22, 1994.

44. A. Trivedi, Strength and dilatancy of jointed rocks with granular fi ll, Acta Geotechnica, Vol. 5, No. 1, pp. 15–31, Aug. 2010.

45. V.K. Arora and A. Trivedi, Effect of Kaolin gouge on strength of jointed rocks, in Asian Regional Symposium on Rock Slopes, 7–11 December 1992, New Delhi, India: proceedings, 1992, pp. 21–26.

46. A. Trivedi and V.K. Arora, Discussion of ‘Bearing Capacity of Shallow Foundations in Anisotropic Non-Hoek–Brown Rock Masses’ by M. Singh and K.S. Rao, J. Geotech. Geoenviron. Eng., Vol. 133, Issue 2, pp. 238–240, 2007.

47. A. Trivedi and V.K. Sud, Settlement of compacted ash fi lls, Geotechnical and Geological Engineering, Vol. 25, No. 2, pp. 163–176, Oct. 2007.


48. A. Trivedi, R. Pathak, and R. Gupta, A common collapse test for gran-ular materials, in: Indian Geotechnical Conference, 2009, pp. 47–51.

49. K. Terzaghi, Theoretical Soil Mechanics. John Wiley & Sons, Inc.USA, 1943.

50. S. Singh, Liquefaction characteristics of silts, Geotechnical and Geological Engineering, Vol. 14, No. 1, pp. 1–19, 1996.

51. A.J. Lutenegger and R.T. Saber, Determination of collapse potential of soils, ASTM Geotechnical Testing Journal, Vol. 11, No. 3, 1988.

52. A.B. Fourie, D. Rowe, and G.E. Blight, The effect of inÆltration on the stability of the slopes of a dry ash dump, Geotechnique, Vol. 49, No. 1, pp. 1–13, 1999.

53. B. Indraratna, P. Nutalaya, K.S. Koo, and N. Kuganenthira, Engineering behaviour of a low carbon, pozzolanic fl y ash and its potential as a con-struction fi ll, Canadian Geotechnical Journal, Vol. 28, No. 4, pp. 542–555, 1991.

54. L. Barden, A. McGown, and K. Collins, The collapse mechanism in partly saturated soil, Engineering Geology, Vol. 7, No. 1, pp. 49–60, 1973.

55. J.H. Schmertmann, Static cone to compute static settlement over sand, J. Soil Mech. Found. Div., ASCE, Vol. 96, No. SM3, pp. 1011–1043, 1970.

56. P.S. Toth, H.T. Chan, and C.B. Cragg, Coal ash as structural fi ll, with special reference to Ontario experience, Canadian Geotechnical Journal, Vol. 25, No. 4, pp. 694–704, 1988.

57. R. Turgeon, Fly ash fi lls a valley, Civil Engineering—ASCE, Vol. 58, No. 12, pp. 67–68, 1988.

58. K. Terzaghi and R.B. Peck, Soil Mechanics in Engineering Practice, Wiley, New York, pp. 64–65, 1948.

59. S.W. Perkins and C.R. Madson, Bearing capacity of shallow founda-tions on sand: A relative density approach, J. Geotech. Geoenviron. Eng., Vol. 126, No. 6, p. 521, 2000.

60. D.J. Dappolonia, E.E. Dappolonia, and R.F. Brissette, Settlement of spread footings on sand, Journal of Soil Mechanics & Foundations Div, Vol. 94, No. SM3, pp. 735–758, 1968.

61. G.G. Meyerhof, Penetration tests and bearing capacity of cohesionless soils, Journal of the Soil Mechanics and Foundation Division, Vol. 82, No. 1, pp. 1–19, 1956.

62. J.B. Burland, M.C. Burbidge, E.J. Wilson, and Terzaghi, Settlement of foundations on sand and gravel, ICE Proceedings, Vol. 78, No. 6, pp. 1325–1381, 1985.